Same old story, still turmoil and uncertainty - banking software

M. Arthur Gillis

SAME OLD STORY, STILL TURMOIL AND UNCERTAINTY

In 1968 it was extremely difficult to procure and implement a complete banking software package for a mainframe.

In 1988 it is still extremely difficult to procure and implement a complete banking software package for a mainframe.

With all the innovative advancements that have occurred in bank automation during the past 20 years, why is the industry still in turmoil? The following examples clearly point out that all is not well with banking software.

* With the support of 28 large and prosperous banks and an investment of more than $60 million, Anacomp, Inc. of Indianapolis, was unable to build and deliver a system in 1983. For whatever reasons one wishes to offer (some say it was the too many cooks in the kitchen syndrome), it is clear that the project did indeed fail.

* Even with a head start that goes back to 1968, the pioneer of bank software, Kirchman Corp., Altamonte Springs, Fla., has been unsuccessful in convincing large banks that its Dimension 4000 product is the system of the '90s.

* The system from Dallas-based Hogan, now marketed by IBM in the U.S., offers the latest technology in architecture, integration and flexibility for customization, but is still not the declared winner. Several false starts, late deliveries, product abortions, staff layoffs, operating losses and corporate redirections have hampered the company. And this is occurring two-and-a-half years after the salvage attempt by the mighty IBM.

* The system from Computer Associates International, Garden City, N.Y., (formerly Uccel) is still beta testing its new all-purpose loan application with targeted deliveries next year.

* The system from Marshall & Ilsley Bank, Milwaukee (marketed by Software Alliance Corp., Berkeley, Calif.), qualifies as a working system with no major failures. But the system has only 50 users, raising questions as to why more banks are not buying the system.

* The system from Systematics, Little Rock, Ark., works for the vendor in a facilities management environment where it is disguised with a full computer service delivery. But in a detailed evaluation of software features, Systematics does not win.

* Cullinet Software Inc., Westwood, Mass., has spent over four years retrofitting a modest banking system (acquired from Bob White Software) into its IDMS database architecture. But the product is still in beta test.

* Unisys, Inc., Blue Bell, Penn., claims to have a home-grown software package for large banks, but it's old and rather dormant.

* NCR, Dayton, Ohio, is a major supplier of computer-related equipment to the banking industry and has the largest customer base of software users in the small bank marketplace. But NCR has not ben able to command a strong presence with its mainframe software.

* When a leading user of technology (BancOne, Columbus, Ohio) hires a leading builder of technology (EDS of Dallas) to create a bank software system, then the obvious question is, Why in 1988 would a bank not opt to buy a ready-made system rather than reinvent the wheel?

While MIS chiefs in all industries would hardly think of building their own general ledger, payroll, fixed assets and accounts payable software, why are several bank MIS chiefs still building their own proprietary bank software?

The above conditions provide a good enough staging to establish the fact that all is not well in the mainframe bank software business. The next step is to determine why.

Hundreds of new computer-related hardware components are announced each day in the U.S., and we are not surprised the they usually work when delivered. Not so in the software business. Even the new Lotus spreadsheet, 1-2-3 version 3.0, is late because it does not work correctly yet. If one were to look at a single reason why bank software is so difficult it would be appropriate to first recognize that software in general is difficult. There are hundreds of specifics that impact bank software. Some of the main ones include:

A. The Everything for Everyone

Dilemma.

Contrary to most outsiders' impressions, bankers are not a bunch of conservative, unimaginative custodians of money. They can dream up some weird combinations of products and customer services that drive MIS chiefs up the wall.

At a product development meeting at Hibernia National Bank in New Orleans, a marketing group invented a new type of deposit account so unique that it was not available anaywhere in the industry. As a matter of fact, it never made it past the legal department because it violated regulatory statutes.

Surprisingly, however, Gary Winker, the MIS director and a Hogan user, was ready to deliver if he had received the go-ahead.

Mike Elvir, executive vice president of Bank Operations at BancOne Services Corp. in Columbus, Ohio, says that his bank is a strong advocate of buying software. They even bought Hogan products and implemented the time deposits system. But in August 1986 BancOne engaged EDS to build a system to handle requirements that are expected in the next decade.

Norwest Bank of Minneapolis enrolled the project in January 1987. Elvir says they examined both Hogan and M&I products and concluded they were good systems but not quite right for a very large, innovative bank. He says there were concerns about throughput and response time, functional flexibility and architectural philosophies.

Curiously, the second and third points sound like quotations from Hogan's sales literature. But Elvir is not convinced that any vendor has the right answers.

He believes that new features and changes to existing features take too long to implement. He says the process needs to be shortened to keep up with the dynamics of a competitive marketplace. To do this job right, he says the architecture has to change from process-driven to information-driven.

He envisions a design that downloads much of the functional aspects at the departmental level with the core processing and customer database at the mainframe, an idea that he admits is based partly on CIF (Customer Information File).

Elvir said he is confident about the EDS project but is still a bit cautious. He says his company will know for sure in two years and after an investment of several million dollars.

But Coley Clark, EDS's top executive for financial institutions, is not going to sit back and wait two years. He sees a major milestone for delivery of the first segment by year-end.

When state-of-the-art software vendors try to build omnifarious bank software they violate the most important rule--build to a firm specification. Instead, they build to an open-ended spec that will cater to whatever a banker invents.

This may be a noble effort to satisfy the customer, but it is a disaster for the people who have to create such a product. It is almost like building a space shuttle today that could be used to someday land on a newly discovered planet that no one knows about.

B. Industry Dynamics Strike at

the Heart of MIS.

Until 1980, banking was a pretty secure and stable business, almost dull. Then deregulation hit and one would think that Congress had it in for the poor MIS guy. Change is very common now in banking and, unfortunately, computer systems do not take kindly to change. Deregulation meant that systems had to change to cater to a wide host of new variations of services. This then caused other trends, such as mergers and profit improvement programs.

Now, with more giant-sized banks serving several unique geographic markets, the systems have to become more powerful and diverse. And finally, because bank profit margins are squeezed, guess who has to tighten his belt and deliver all this new stuff at an affordable price?

These dynamics again point to one more reason why software developers were having trouble responding to the needs. It was indeed a moving target.

C. Employee Turnover is Like Losing

the Combination to the Vault.

Employee longevity surveys show that the computer industry is notoriously unstable, especially in the software vendor sector.

Who would want to work for a software vendor anyway? The pressures are enormous and although salaries are high, when computed on the basis of hours worked, the real pay may approach minimum wage.

Developers of business systems in general believe that genuis in software design and programming is relatively low on the list of key qualifications. Continuity and stability are more important.

But in the bank software business it is doubtful if any of the original creators of currently marketed software are still around. Picking up someone else's pieces causes enormous delays and is one more reason for today's problems.

D. Mainframe Software Developers

Start with Two Strikes

Against Them Just Beause of the

Mainframe.

In fairness to MIS types in other industries, it should be pointed out that banking applications are not so complex to require genius-type developers. Other industry systems are much more complex.

For example, if EDS can successfully automate the Saginaw Vanguard plant for General Motors, surely they can build a banking system for BancOne and Norwest. Whether these banks can spend as much as GM is another matter. The reason for constant delays in product development may be due not to the software--but to the hardware.

Most developers agree that mainframe software is a lot tougher to build than, for example, mini-based software. The following is proof: There are 48 companies that have built a bank software system that runs on minis including IBM, NCR, Unisys, Texas Instruments, DEC, DG and Wang. That is such a huge number one has to believe it was a no-brainer for many of these companies to build their product.

From those 48, there are 14 superior choices, and they represent approximately 90% of the industry's installed base of small banks. The capabilities of these systems are as strong, if not better, than those of some mainframe software systems. So one can argue that if 14 small companies can build good software on minis, why cannot eight large companies build good software on mainframes?

Here is another recent development that further supports the above hypothesis.

Three people working 50 hours per week for one year have just completed a modern integrated bank system that will run on the IBM AS/400. As viewed through the eyes of a discriminating banker, the system will look as good--if not better--than the best mainframe system.

E. You Can't Please Stockholders

and Customers at the Same Time.

Time is running out. While banks are demanding more from their software vendors, stockholders are looking for profits that are long overdue. The pursuit of the ultimate bank software package is not a new endeavor. It got its real start about eight years ago.

Consider the following comparison. Systematics is characterized as a tightward spender but a provider of reliable computer services. Their customers are typically not industry innovators.

Systematics is like the ice cream store that only serves vanilla, chocolate and strawberry, instead of the store that advertises 35 flavors but never seems to have the one you want when you want it.

Yet Systematics is a consistent performer in terms of earnings and is a favorite of Wall Street.

Hogan Systems, on the other hand, has a perceived image as the leader in technology. It also has a customer base to prove it, comprising most of the movers and shakers. But Hogan has not had a consistent earnings record since the company started 10 years ago.

This kind of pressure to perform is not healthy, and it means that even though there are only eight vendors to choose from right now, the list may be shortened next year.

Have no fear. Banks may fail because of bad loans to Third World countries, farmers, real estate speculators and oil patch businesses. But they will not fail because of a marginal or even weak automated system. There are viable solutions, but in every case one has to be realistic and accept some concessions.

Here is a simple but practical offering of ideas to explore.

1. Good Plain Vanilla Software.

If a bank is willing to forego some of the sizzle of state-of-the-art technology, there are three good vendors that offer complete systems for IBM mainframes. They are Systematics, Kirchman Corp. and Computer Associates International.

Ray Maturi, head of marketing for Systematics, says the company conducts a step-by-step development approach that evolves into a working product.

He is also critical of those firms that jump in with half-baked but glamorous solutions that do not produce a steady stream of profits for the company.

Bahram Yusefzadeh, head of the Product and Marketing Strategies Division for the Kirchman Corp., does not take kindly to the reference of plain vanilla software. He points to an existing customer base of 180 financial institutions.

When asked how he plans to outdo the hi-tech competition, eh calmly said that the company's Dimension 4000 product is complete. He said it works and it is easy to implement.

John Capabianco, director of product marketing and banking applications for CAI, says their philosophy in product announcements comes directly from the man in charge, Charles Wang.

Capabianco, echoing a wine commercial, said CAI "will not release any product before its time."

This was not always the case prior to the acquisition of Uccel. Capabianco admits that it is easier to sell promises and "trust me's" than a ready-to-deliver product. Bankers will buy an image easily, but they will severely criticize something they can sink their teeth into. Wang's philosophy is, "Don't sell what bankers want to hear. Sell them what you have and what they need today."

Julian Beard, executive VP at First Security National Bank in Lexington, Ky., says, "We'd love to have the benefits of integrated software but not at today's prices. So we operate effectively with a combination of Uccel, Kirchman, MSA and AFS."

2. Closer to Today's Technology.

Two vendors currently share the spotlight for integrated software that seems to be in vogue among forward-thinking bankers. They are Hogan Systems, which is sold by IBM, and M&I, sold by Software Alliance Corp. Brian Scott, president of Banks of Iowa Computer Services (BICS), Cedar Rapids, Iowa, bought the Hogan deposits system in 1980. At the same time he took some of the responsibility of managing the software.

It was like a partnership then because they were skeptical about Hogan's support and they wanted to be self-sufficient. Today, Scott says it is "the mostsolid piece of software they have ever bought. It has excellent reliability, good architecture, is highly parameter-driven and runs efficiently."

But Bics is not an ordinary passive user. It employs some brain trusts that know how to improve software vendor offerings.

Fred Cisewski, director of MIS at BankSouth in Atlanta, says, "We're halfway into implementation of the whole system and we're meeting expectations. With IBM as a Hogan partner, we know we're in the right arena."

Gary Fiedler, chairman and CEO at Hogan Systems, is no newcomer at Hogan, even though he took charge this past year. He also was a former customer and has always been close to the company as a director.

"The company's new direction is to focus on what customers need and want," Fiedler said. "Most vendor brochures focus on the vendor. Hogan likes to talk about customers." (Hogan customers do talk about their vendor as a partner.)

A few years ago when Hogan experienced financial problems, several bankers who were interviewed about the problem saw it as a non-event. Their answer was, "If they need money we'll supply enough to keep them on the job. We want the product and we'll pay for it one way or the other."

Cyril Reif, product manager, finance, at IBM in Atlanta, says IBM is very bullish on Integrated Banking Applications (IBA), which is the IBM version of the Hogan system.

He admits, however, that it was a long learning curve with many complex issues. But IBM is now on top of the problem as evidenced by new sales made by IBM is now on top of of Hogan customers to IBA.

Joseph M. Guglielmi, IBM president of the Application Systems Division in Milford, Conn., Stated, "The learning process is now behind us and we expect our relationship will bear fruit."

Jon Blankmeyer, president of Software Alliance Corp., Berkeley, Calif., does not sound like the euphoric vendor of magical software. The track record is clean, and he says the company is not compelled to sell volume, it wants to sell good business. He believes market size expectations are overblown. He sees a very limited number of buyers. With 50 customers, he is targeting growth from the existing base with new products. One of these is a migration to DB2 from DL1.

James Boland, senior vice president of Information Services at Deposit Guaranty National Bank in Jackson, Miss., is an M&I user, and he confirms Blamkeyer's statement with greater conviction.

Boland says he could not be more pleased with the customer module and the deposits module. "M&I's support is outstanding and we are very pleased with what we have. But we didn't buy the loan module because we don't believe it's ready yet. M&I will get there in two years and then we'll buy."

Philip Cunningham, senior vice president of BancOklahoma Corp., Tulsa, Okla., says, "We're very interested in the single database concept of customer information. That means we regard integration as a critical architecture. We've taken the first step with the M&I deposits module but we stopped because in Oklahoma banks don't spend money anymore. So our systems are taped together using Uccel's CIF and Shaw Systems Associates loan system just to get the job done."

3. Future Possibilities.

Cullinet may be the answer for users who are in no hurry for new product offerings. Jim Fox, vice president of banking products at Cullinet, says a new system will be released from a three-bank beta test by year-end. It will be positioned technologically to address the needs of the '90s.

For expects Cullinet will overcome its image of a latecomer and as an unfamiliar vendor in bank software by introducing an advanced product that will outpace the competition. Those are strong promises in an industry that has been burned by similar claims in the past. Already Cullinet has aborted one piece of its penetration into the banking industry; it no longer plans to market a banking system for DEC computers.

Cullinet will have a double challenge in selling to an audience of skeptics, in that they'll have to sell IDMS as well as the banking system.

One could speculate that after the EDS system is delivered to BancOne and Norwest it could reach the marketplace as a vendor-supported product. EDS does not just "sell a tape," however, as Coley Clark puts it. EDS will support it with value-added services.

4. Unisys Users.

Because large banks are primarily IBM users, little third-party software is available for Unisys or NCR mainframes. Unisys has its own system called Global Financial System. In addition, there are some turnkey vendors who are climbing up from serving small banks to going after the big banks. They are Information Technology, Inc., Lincoln, Neb.; Newtrend Group, Orlando, Fla.; and Jack Henry & Associates, Monett, Mo.

These three are excellent companies in their own league, but can they hit home runs in the big leagues? Or better yet, will the big guys pitch enough balls to them to try for the homeruns? Big bank MIS types are very sensitive about stats and pedigree. They think of mini-based software vendors as little leaguers. Look again.

Don Dillon, president of Information Technology, Inc., has in 10 years, built the banking industry's most successful software company in terms of performance and customer satisfaction. His philosophy is that one should start first with a good base and build onto it in time. (The big guys tried to build the ultimate system immediately and failed. Some tasks are just too global and difficult, and throwing lots of bodies at the project just makes it more difficult to control.)

NCR has been a sleeper in the large bank arena. But by year-end a new product called Universal Financial System will come out of beta test. Peter Augusta, assistant vice president of Financial Systems, says, "We're planning an all-out effort to go after the big bank market. We have 23 contracts now for UFS."

5. A Little Bit of Potpourri.

There was indeed life prior to integrated bank software. It usually consisted of individual applications tied together with software bridges. As a result, some banks are satisfied that having something that works is a lot better than risking large conversions to the world of uncertainty.

First Wisconsin National Bank, which services its own banks as well as correspondent banks equal to $20 billion in assets, is such a bank. Gerald Ried, first vice president, says, "We built all our own software and are very pleased with the results. In a couple of cases we found better software than we could build, so we bought Uccel's CIF and Shaw Systems Associates commercial loan application."

Henry Garcia, senior vice president of Systems Development at Citizens and Southern Corporation in Atlanta, a $22 billion holding company, says his bank did an intensive study one year ago and concluded there wasn't one single answer.

"We bought the IBM version of the Hogan system for Customer Information System (CIS) only and then we bought other vendor packages for general ledger, trust, credit card and installment loans. Our eventual system will consist of several brand names but within the holding company (Georgia, South Carolina and Florida) our banks will be running on C&S common systems," he said.

6. The No Software Solution.

Although it is not a sweeping trend, some banks elect to solve their software dilemma--and their entire data processing dilemma--by switching from in-house systems to a service bureau.

George Dalton, chairman of FIserv, Milwaukee, a computer services firm for financial institutions, says, "We're in the software business also. We don't sell it as a product. We use it as the main resource in delivering quality solutions.

"But we know how to manage our own software. If we sold it to someone else, we're not sure they could handle it as well, and they would be leaning on us for support. That's a business we don't want to be in, thank you," he said.

SUBTLE DIFFERENCES

The U.S. banking industry consists of 1,664 thrifts and banks that are over $250 million in assets and thus qualify as mainframe users.

But since many of them use a service bureau or facilities management contract, or squeeze by on a supermini, the true buyers of mainframe software to run their own MIS division add up to only about 500. The makeup and culture of the 500 vary substantially but not from the lobby view. One has to look in the board room and the back room to notice differences. Here are the categories:

* The 20 at the very top belong to an elite group that builds its own technology. For example, Citibank built its own ATMs. They generally do not buy software.

* Another group consists of banks that are very pragmatic and they buy systems that work even if they do not have all the bells and whistles. They will not spend dollars for leading edge because they do traditional banking well and basic software works for them.

* The third group consists of banks that have built in-house software back in the 1960s because that was the only way, and they still live with that same software.

A common trait within this group is that they are located in certain regions of the country where people accept one job for life. The same people who created the software enhance it today. When they retire in the next 10 years, this group will become buyers of software.

* The fourth group consists of banks that are very creative with unusually innovative needs. It is this group that will never be completely satisfied with today's systems. They want tomorrow's systems and are the toughest customers to please. And this is the group that contributes most to the turmoil and uncertainty that characterizes mainframe-based banking software.

COPYRIGHT 1988 Wiesner Publications, Inc.
COPYRIGHT 2004 Gale Group